Turning program spreadsheets into an honest funder impact report
Every grant comes with a report, and in a small org the data lives in spreadsheets, sign-in sheets, and the ED's memory. Nonprofits without an analyst report that AI-assisted analysis of program data is a step change — the job is turning raw numbers into a narrative mapped to the funder's objectives without fudging anything.
You are a program evaluator and report writer for a small nonprofit. Help me draft a grant report for a funder. The grant's stated objectives and targets (from our grant agreement): {{grant_objectives}} Our actual data for the reporting period (pasted from spreadsheets, may be messy): {{program_data}} Reporting period and funder: {{reporting_details}} Do this in two parts. PART 1 — ANALYSIS: For each grant objective, compare our actual results to the target. State plainly: met, exceeded, or missed, with the numbers. Note any trends across the period (growth, drop-offs, seasonal patterns) that the data actually supports. PART 2 — DRAFT REPORT: Write narrative sections mapped one-to-one to the grant objectives. Where we missed a target, use honest framing: what happened, what we learned, what we're changing — no spin, no excuses. Include a "looking ahead" paragraph grounded only in plans I've stated. Hard rules: use only the data I provided — never estimate, extrapolate, or fill gaps with plausible numbers. Where data is missing for an objective, write [DATA NEEDED: description]. If my numbers appear inconsistent, list the discrepancies before writing anything.
Fill in your details and the prompt updates live — then copy.
**Objective 2 — Distribute 250,000 lbs of food: EXCEEDED.** Actual distribution reached 268,400 lbs, 107% of target, with quarterly volume growing steadily (Q1 to Q3). **Objective 3 — Establish 3 new pantry partnerships: MISSED (2 of 3).** Grace Church and Northside Elementary partnerships launched in February and May. A third prospective partner paused talks over storage capacity. We are addressing this by [DATA NEEDED: current status of third-partner pipeline and revised timeline]. Note: unduplicated households (389) sits just below the 400 target with one quarter remaining.
The full workflow
- Clean obvious errors from your spreadsheet before pasting; strip client names and identifying details first.
- Run Part 1 and confirm every comparison against your own records before drafting narrative.
- Fill each [DATA NEEDED] flag from real records — or tell the funder the data isn't collected yet.
- Add one concrete story (with consent) that the numbers can't show; that's the part only you can write.
- Keep the final report and underlying data together in your grant file for the audit trail.
Watch out for
Never let AI fill data gaps with plausible-sounding numbers. Misstating results to a funder can end the relationship, and false statements on federal grant reports create False Claims Act liability.
Client and beneficiary records are confidential — remove names, addresses, and case details before pasting service data into any AI tool, and use aggregate numbers wherever possible.
Honest missed-target framing usually strengthens funder trust; AI's instinct to spin is a bug, not a feature.
Where this comes from
Every use case on this site is grounded in real reports from working nonprofit directors — not invented by us.